Streamlined Toolkit for Real-time Exploratory Analysis of Multiomics

Streamlined Toolkit for Real-time Exploratory Analysis of Multiomics

Decoding RNA expression matrix…
GSEA
GSVA
GSEA
GSVA
GSEA
GSVA
GSEA
GSVA

Squares represent transcription factors, circles are target genes. Node colour reflects log fold change (red = higher in the test group, blue = higher in the reference group). Green edges indicate positive influence; red edges indicate negative influence. Squares represent transcription factors, circles are target genes. Node colour reflects log fold change using a purple–orange scale (purple = higher in the test group, orange = higher in the reference group). Teal edges indicate positive influence; orange edges indicate negative influence. Hovering highlights neighbours and clicking a node updates the expression box plot on the right.

Loading...

Integrated tissue overview

Detailed narrative

Per-celltype quick card

Alerts

Key findings

Composition shifts

Tissue-level genes

Top upregulated

    Top downregulated

      Pathway programs

      Candidate regulators

      Cell–cell interactions

      Disease links & therapeutics

      Hypotheses & validation ideas

      Robustness & anomalies

      Run metadata

      Methods — STREAM v2.0 dashboard build

      Data inputs & preprocessing
      Per-cell-state pseudobulk differential expression
      Gene-set, pathway, and regulator scoring
      Expression payload & rendering
      AI insights — optional

      Enabling “Generate biology insights” triggers LLM-backed summaries per cell type. The model receives only aggregate artefacts (cluster sizes, edgeR up/down lists, top BTM/Hallmark/TF hits and their scores) together with study context from PROJECT_TITLE/PROJECT_CONTEXT/CONTRAST_VAR. Outputs are harmonised into a strict Markdown template with canonical/fine labels, confidence, and key markers, and an overview prompt synthesises tissue-wide takeaways. Raw counts or per-cell matrices never leave the session.

      Dashboard tabs — content, statistics, AI usage
      Show general AI prompt

      QC and composition tabs reflect the layer chosen at run time; when multiple layers are available the same dashboard can be regenerated for alternative normalisations.

      STREAM

      Streamlined Toolkit for Real-time Exploratory Analysis of Multiomics

      Interactive data mining and analysis of bulk and single-cell expression data — delivered as a fast, shareable HTML dashboard with UMAPs, marker discovery, and enrichment for BTM modules and Hallmark pathways.

      About the toolkit

      STREAM is a zero-install workflow that transforms differential-expression analyses from bulk, single-cell and pseudobulk RNA-seq experiments into an interactive HTML dashboard. It is designed to make exploratory analysis reproducible and accessible by automating quality control (QC), clustering, marker discovery, enrichment analysis and multi-scale summarization. Instead of requiring collaborators to rerun notebooks, STREAM bundles all analyses, including interactive plots, statistics and provenance, into a single offline-capable web page.

      About this dataset
      Dataset
      COVID
      Layer
      counts
      Technology
      10X V3
      Cells
      Median UMIs / cell
      Total UMIs
      Median genes / cell
      Samples
      Generated
      2025-10-21T11:47:11
      Version
      2.0
      Notes: values shown reflect the current HTML build; analyses respect the selected layer.
      Contributors
      Support & Acknowledgements

      Single Cell MultiOmics Lab

      If you use STREAM in your work, please cite: STREAM (v2.0) — Streamlined Toolkit for Real-time Exploratory Analysis of Multiomics, generated 2025-10-21T11:47:11.

      Legal & Licensing

      Disclaimer. STREAM is a research and education tool. It is not intended for clinical, diagnostic, or patient-management decisions.

      • AI Insights (if enabled) are LLM-generated summaries that can be incomplete or incorrect; always validate with markers, BTM modules, and Hallmark pathways.
      • Privacy. This dashboard does not upload sample-level data to a server; rendering happens in your browser. (If you need strict offline use, bundle local JS/CSS assets instead of CDNs.)
      • Upstream AI usage. If you tick “Generate biology insights” during build, cluster-level summaries (markers, enrichment ranks, etc.) are sent to the configured AI provider to produce text; raw counts are not sent.
      • No warranty. Provided “AS IS” without warranty of any kind; you are responsible for verifying results and for compliance with any data-use terms.
      STREAM code license: MIT License
      Dataset licence: CC BY 4.0 (Example – replace per dataset)
      Show full STREAM license